In a strategic move to enhance its autonomy in artificial intelligence systems, OpenAI has appointed Peter Steinberger, the founder of the innovative open-source project OpenClaw. This recruitment aligns with the company’s goal to expand its capabilities and maintain a competitive edge against rivals like Anthropic and Google.
OpenClaw, which has garnered significant attention for enabling users to create personalized AI agents, allows these agents to manage various tasks across users’ digital lives. Formerly known as Clawdbot and Moltbot, the project gained traction particularly over the past month as users began deploying the software on their own devices, incorporating it into platforms such as WhatsApp, Slack, and iMessage. This functionality enables the agents to manage emails, calendars, and other aspects of a person’s daily routine, acting autonomously on their behalf.
Sam Altman, OpenAI’s CEO, highlighted Steinberger’s brilliance and innovative approach in a post on social media platform X, expressing optimism that his contributions will be integral to the development of personal AI agents. “Peter Steinberger is joining OpenAI to drive the next generation of personal agents,” said Altman, noting Steinberger’s unique vision for intelligent agents that interact with each other to effectively serve users.
As the agent platform grows, it has sparked intriguing interactions. Agents created on OpenClaw can communicate with one another, even discussing creations such as a fictional religion called “Crustafarianism.” This has led to debates among both the AI agents and human observers regarding the extent of understanding these systems possess about their discussions and creations.
Steinberger, an Austrian developer, initially prototyped OpenClaw in a mere hour, and by early February, the platform had amassed around 1.5 million agents. He revealed in a recent podcast that maintaining the project incurs monthly costs between $10,000 and $20,000, adding that despite these expenses, he currently operates at a loss.
Both OpenAI and Steinberger confirmed that OpenClaw will retain its independence as a foundation while continuing to benefit from OpenAI’s support. Altman emphasized the importance of maintaining an open-source approach, stating, “The future is going to be extremely multi-agent and it’s important to us to support open source as part of that.”
The project, inspired by the lobster, underwent a name change to avoid confusion with Anthropic’s Claude chatbot, which led to a request for rebranding. Anthropic itself has developed agents, evidenced by the launch of Claude Cowork earlier this year, aimed at simplifying automation tasks for non-technical users.
However, as the capabilities of such AI agents grow, security experts have raised alarms over potential risks following their access to sensitive data, including financial information. OpenAI intends to leverage Steinberger’s expertise to further develop its agent-based offerings, including its coding platform, Codex.
In a blog post announcing his transition to OpenAI, Steinberger shared his motivations for his work, expressing a desire to make AI accessible and fun for broader audiences. “What I want is to change the world, not build a large company,” he declared, indicating that his partnership with OpenAI represents the most efficient pathway to realize this ambition. He added, “My next mission is to build an agent that even my mum can use, which will require significant changes and careful consideration for safety.”
